Apply Logistic Regression¶
In [1]:
%matplotlib inline
import warnings
warnings.filterwarnings("ignore")
import sqlite3
import pandas as pd
import numpy as np
import nltk
import string
import matplotlib.pyplot as plt
import seaborn as sns
from sklearn.feature_extraction.text import TfidfTransformer
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.feature_extraction.text import CountVectorizer
from sklearn.metrics import confusion_matrix
from sklearn import metrics
from sklearn.metrics import roc_curve, auc
from nltk.stem.porter import PorterStemmer
import re
# Tutorial about Python regular expressions: https://pymotw.com/2/re/
import string
from nltk.corpus import stopwords
from nltk.stem import PorterStemmer
from nltk.stem.wordnet import WordNetLemmatizer
from gensim.models import Word2Vec
from gensim.models import KeyedVectors
import pickle
from tqdm import tqdm
import os
from chart_studio.plotly import plotly
import plotly.offline as offline
import plotly.graph_objs as go
offline.init_notebook_mode()
from collections import Counter
1) Splitting data into Train and cross validation(or test): Stratified Sampling¶
In [2]:
project_data = pd.read_csv("train_data.csv", nrows = 50000)
resource_data = pd.read_csv("resources.csv", nrows = 50000)
In [3]:
print("Number of data points in train data", project_data.shape)
print('-'*50)
print("The attributes of data :", project_data.columns.values)
In [4]:
# Let's check for any "null" or "missing" values
project_data.info()
In [5]:
project_data['teacher_prefix'].isna().sum()
Out[5]:
In [6]:
# "teacher_prefix" seems to contain 3 "missing" values, let't use mode replacement strategy to fill those missing values
project_data['teacher_prefix'].mode()
Out[6]:
In [7]:
# Let's replace the missing values with "Mrs." , as it is the mode of the "teacher_prefix"
project_data['teacher_prefix'] = project_data['teacher_prefix'].fillna('Mrs.')
In [8]:
project_data['teacher_prefix'].value_counts()
Out[8]:
In [9]:
price_data = resource_data.groupby('id').agg({'price':'sum', 'quantity':'sum'}).reset_index()
project_data = pd.merge(project_data, price_data, on='id', how='left')
In [10]:
# Let's select only the selected features or columns, dropping "project_resource_summary" as it is optional
#
project_data.drop(['id','teacher_id','project_submitted_datetime','project_resource_summary'],axis=1, inplace=True)
project_data.columns
Out[10]:
In [11]:
# Data seems to be highly imbalanced since the ratio of "class 1" to "class 0" is nearly 5.5
project_data['project_is_approved'].value_counts()
Out[11]:
In [12]:
number_of_approved = project_data['project_is_approved'][project_data['project_is_approved'] == 1].count()
number_of_not_approved = project_data['project_is_approved'][project_data['project_is_approved'] == 0].count()
print("Ratio of Project approved to Not approved is:", number_of_approved/number_of_not_approved)
Let's first merge all the project_essays into single columns
In [13]:
# merge two column text dataframe:
project_data["essay"] = project_data["project_essay_1"].map(str) +\
project_data["project_essay_2"].map(str) + \
project_data["project_essay_3"].map(str) + \
project_data["project_essay_4"].map(str)
In [14]:
project_data.head(2)
Out[14]:
In [15]:
# Let's drop the project essay columns from the dadaset now, as we have captured the essay text data into single "essay" column
project_data.drop(['project_essay_1','project_essay_2','project_essay_3','project_essay_4'],axis=1, inplace=True)
In [16]:
y = project_data['project_is_approved'].values
X = project_data.drop(['project_is_approved'], axis=1)
X.head(1)
Out[16]:
In [17]:
# train test split
from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.33, stratify=y)
2) Make Data Model Ready: encoding numerical, categorical features¶
In [18]:
def cleaning_text_data(list_text_feature,df,old_col_name,new_col_name):
# remove special characters from list of strings python: https://stackoverflow.com/a/47301924/4084039
# https://www.geeksforgeeks.org/removing-stop-words-nltk-python/
# https://stackoverflow.com/questions/23669024/how-to-strip-a-specific-word-from-a-string
# https://stackoverflow.com/questions/8270092/remove-all-whitespace-in-a-string-in-python
feature_list = []
for i in list_text_feature:
temp = ""
# consider we have text like this "Math & Science, Warmth, Care & Hunger"
for j in i.split(','): # it will split it in three parts ["Math & Science", "Warmth", "Care & Hunger"]
if 'The' in j.split(): # this will split each of the catogory based on space "Math & Science"=> "Math","&", "Science"
j=j.replace('The','') # if we have the words "The" we are going to replace it with ''(i.e removing 'The')
j = j.replace(' ','') # we are placeing all the ' '(space) with ''(empty) ex:"Math & Science"=>"Math&Science"
temp+=j.strip()+" " #" abc ".strip() will return "abc", remove the trailing spaces
temp = temp.replace('&','_') # we are replacing the & value into
feature_list.append(temp.strip())
df[new_col_name] = feature_list
df.drop([old_col_name], axis=1, inplace=True)
from collections import Counter
my_counter = Counter()
for word in df[new_col_name].values:
my_counter.update(word.split())
feature_dict = dict(my_counter)
sorted_feature_dict = dict(sorted(feature_dict.items(), key=lambda kv: kv[1]))
return sorted_feature_dict
In [19]:
def clean_project_grade(list_text_feature,df,old_col_name,new_col_name):
# remove special characters from list of strings python: https://stackoverflow.com/a/47301924/4084039
# https://www.geeksforgeeks.org/removing-stop-words-nltk-python/
# https://stackoverflow.com/questions/23669024/how-to-strip-a-specific-word-from-a-string
# https://stackoverflow.com/questions/8270092/remove-all-whitespace-in-a-string-in-python
feature_list = []
for i in list_text_feature:
temp = i.split(' ')
last_dig = temp[-1].split('-')
fin = [temp[0]]
fin.extend(last_dig)
feature = '_'.join(fin)
feature_list.append(feature.strip())
df[new_col_name] = feature_list
df.drop([old_col_name], axis=1, inplace=True)
from collections import Counter
my_counter = Counter()
for word in df[new_col_name].values:
my_counter.update(word.split())
feature_dict = dict(my_counter)
sorted_feature_dict = dict(sorted(feature_dict.items(), key=lambda kv: kv[1]))
return sorted_feature_dict
2.1) Text Preprocessing: project_subject_categories¶
In [20]:
x_train_sorted_category_dict = cleaning_text_data(X_train['project_subject_categories'],X_train,'project_subject_categories','clean_categories')
x_test_sorted_category_dict = cleaning_text_data(X_test['project_subject_categories'],X_test,'project_subject_categories','clean_categories')
2.2) Text Preprocessing : project_subject_subcategories¶
In [21]:
x_train_sorted_subcategories = cleaning_text_data(X_train['project_subject_subcategories'],X_train,'project_subject_subcategories','clean_subcategories')
x_test_sorted_subcategories = cleaning_text_data(X_test['project_subject_subcategories'],X_test,'project_subject_subcategories','clean_subcategories')
2.3) Text Preprocessing: project_grade_category¶
In [22]:
x_train_sorted_grade = clean_project_grade(X_train['project_grade_category'],X_train,'project_grade_category','clean_grade')
x_test_sorted_grade = clean_project_grade(X_test['project_grade_category'],X_test,'project_grade_category','clean_grade')
2.4) Text Preprocessing (stowords): project_essay, project_title¶
In [23]:
# https://stackoverflow.com/a/47091490/4084039
import re
def decontracted(phrase):
# specific
phrase = re.sub(r"won't", "will not", phrase)
phrase = re.sub(r"can\'t", "can not", phrase)
# general
phrase = re.sub(r"n\'t", " not", phrase)
phrase = re.sub(r"\'re", " are", phrase)
phrase = re.sub(r"\'s", " is", phrase)
phrase = re.sub(r"\'d", " would", phrase)
phrase = re.sub(r"\'ll", " will", phrase)
phrase = re.sub(r"\'t", " not", phrase)
phrase = re.sub(r"\'ve", " have", phrase)
phrase = re.sub(r"\'m", " am", phrase)
return phrase
In [24]:
# https://gist.github.com/sebleier/554280
# we are removing the words from the stop words list: 'no', 'nor', 'not'
stopwords= ['i', 'me', 'my', 'myself', 'we', 'our', 'ours', 'ourselves', 'you', "you're", "you've",\
"you'll", "you'd", 'your', 'yours', 'yourself', 'yourselves', 'he', 'him', 'his', 'himself', \
'she', "she's", 'her', 'hers', 'herself', 'it', "it's", 'its', 'itself', 'they', 'them', 'their',\
'theirs', 'themselves', 'what', 'which', 'who', 'whom', 'this', 'that', "that'll", 'these', 'those', \
'am', 'is', 'are', 'was', 'were', 'be', 'been', 'being', 'have', 'has', 'had', 'having', 'do', 'does', \
'did', 'doing', 'a', 'an', 'the', 'and', 'but', 'if', 'or', 'because', 'as', 'until', 'while', 'of', \
'at', 'by', 'for', 'with', 'about', 'against', 'between', 'into', 'through', 'during', 'before', 'after',\
'above', 'below', 'to', 'from', 'up', 'down', 'in', 'out', 'on', 'off', 'over', 'under', 'again', 'further',\
'then', 'once', 'here', 'there', 'when', 'where', 'why', 'how', 'all', 'any', 'both', 'each', 'few', 'more',\
'most', 'other', 'some', 'such', 'only', 'own', 'same', 'so', 'than', 'too', 'very', \
's', 't', 'can', 'will', 'just', 'don', "don't", 'should', "should've", 'now', 'd', 'll', 'm', 'o', 're', \
've', 'y', 'ain', 'aren', "aren't", 'couldn', "couldn't", 'didn', "didn't", 'doesn', "doesn't", 'hadn',\
"hadn't", 'hasn', "hasn't", 'haven', "haven't", 'isn', "isn't", 'ma', 'mightn', "mightn't", 'mustn',\
"mustn't", 'needn', "needn't", 'shan', "shan't", 'shouldn', "shouldn't", 'wasn', "wasn't", 'weren', "weren't", \
'won', "won't", 'wouldn', "wouldn't"]
In [25]:
# Combining all the above stundents
from tqdm import tqdm
def process_text(df,col_name):
preprocessed_feature = []
# tqdm is for printing the status bar
for sentance in tqdm(df[col_name].values):
sent = decontracted(sentance)
sent = sent.replace('\\r', ' ')
sent = sent.replace('\\"', ' ')
sent = sent.replace('\\n', ' ')
sent = re.sub('[^A-Za-z0-9]+', ' ', sent)
# https://gist.github.com/sebleier/554280
sent = ' '.join(e for e in sent.split() if e.lower() not in stopwords)
preprocessed_feature.append(sent.lower().strip())
return preprocessed_feature
In [26]:
x_train_essay_preprocessed = process_text(X_train,'essay')
x_test_essay_preprocessed = process_text(X_test,'essay')
In [27]:
x_train_title_preprocessed = process_text(X_train,'project_title')
x_test_title_preprocessed = process_text(X_test,'project_title')
2.5) Vectorizing Categorical Data¶
project_subject_categories (clean_categories)¶
In [28]:
# we use count vectorizer to convert the values into one
from sklearn.feature_extraction.text import CountVectorizer
def cat_vectorizer(X_train,df,col_name):
vectorizer = CountVectorizer()
vectorizer.fit(X_train[col_name].values)
feature_one_hot = vectorizer.transform(df[col_name].values)
print(vectorizer.get_feature_names())
return feature_one_hot, vectorizer.get_feature_names()
In [29]:
x_train_cat_one_hot, x_train_cat_feat_list = cat_vectorizer(X_train,X_train,'clean_categories')
x_test_cat_one_hot, x_test_cat_feat_list = cat_vectorizer(X_train,X_test,'clean_categories')
In [30]:
# shape after categorical one hot encoding
print(x_train_cat_one_hot.shape)
print(x_test_cat_one_hot.shape)
project_subject_subcategory (clean_subcategory)¶
In [31]:
x_train_subcat_one_hot, x_train_subcat_feat_list = cat_vectorizer(X_train,X_train,'clean_subcategories')
x_test_subcat_one_hot, x_test_subcat_feat_list = cat_vectorizer(X_train,X_test,'clean_subcategories')
In [32]:
# shape after categorical one hot encoding
print(x_train_subcat_one_hot.shape)
print(x_test_subcat_one_hot.shape)
school_state¶
In [33]:
# we use count vectorizer to convert the values into one hot encoding
# CountVectorizer for "school_state"
x_train_state_one_hot, x_train_state_feat_list = cat_vectorizer(X_train,X_train,'school_state')
x_test_state_one_hot, x_test_state_feat_list = cat_vectorizer(X_train,X_test,'school_state')
In [34]:
# shape after categorical one hot encoding
print(x_train_state_one_hot.shape)
print(x_test_state_one_hot.shape)
teacher_prefix¶
In [35]:
# we use count vectorizer to convert the values into one hot encoding
# CountVectorizer for teacher_prefix
x_train_teacher_prefix_one_hot,x_train_teacher_prefix_feat_list = cat_vectorizer(X_train,X_train,'teacher_prefix')
x_test_teacher_prefix_one_hot,x_test_teacher_prefix_feat_list = cat_vectorizer(X_train,X_test,'teacher_prefix')
In [36]:
# shape after categorical one hot encoding
print(x_train_teacher_prefix_one_hot.shape)
print(x_test_teacher_prefix_one_hot.shape)
project_grade_category¶
In [37]:
# using count vectorizer for one-hot encoding of project_grade_category
x_train_grade_one_hot, x_train_grade_feat_list = cat_vectorizer(X_train,X_train,'clean_grade')
x_test_grade_one_hot, x_test_grade_feat_list = cat_vectorizer(X_train,X_test,'clean_grade')
In [38]:
# shape after categorical one hot encoding
print(x_train_grade_one_hot.shape)
print(x_test_grade_one_hot.shape)
2.6) Vectorizing Text Data¶
2.6.1) Bag of Words (essay)¶
In [39]:
# We are considering only the words which appeared in at least 10 documents(rows or projects).
def bow_vectorizer(X_train,col_name,df):
vectorizer = CountVectorizer(min_df=10,ngram_range=(2,2), max_features=5000)
vectorizer.fit(X_train[col_name].values)
df_bow = vectorizer.transform(df[col_name].values)
return df_bow, vectorizer.get_feature_names()
In [40]:
x_train_essay_bow, x_train_essay_feat = bow_vectorizer(X_train,'essay',X_train)
x_test_essay_bow, x_test_essay_feat = bow_vectorizer(X_train,'essay',X_test)
In [41]:
print(x_train_essay_bow.shape)
print(x_test_essay_bow.shape)
2.7.2) Bag of Words (title)¶
In [42]:
def bow_vectorizer_title(X_train,col_name,df):
vectorizer = CountVectorizer()
vectorizer.fit(X_train[col_name].values)
df_bow = vectorizer.transform(df[col_name].values)
return df_bow, vectorizer.get_feature_names()
In [43]:
x_train_title_bow, x_train_title_feat = bow_vectorizer_title(X_train,'project_title',X_train)
x_test_title_bow, x_test_title_feat = bow_vectorizer_title(X_train,'project_title',X_test)
In [44]:
print(x_train_title_bow.shape)
print(x_test_title_bow.shape)
2.7.3) TFIDF (essay)¶
In [45]:
from sklearn.feature_extraction.text import TfidfVectorizer
# We are considering only the words which appeared in at least 10 documents(rows or projects).
def tfidf_vectorizer(X_train,col_name,df):
vectorizer = TfidfVectorizer(min_df=10, ngram_range = (2,2), max_features = 5000)
vectorizer.fit(X_train[col_name].values)
df_tfidf = vectorizer.transform(df[col_name].values)
return df_tfidf, vectorizer.get_feature_names()
In [46]:
# Lets vectorize essay
x_train_essay_tfidf, x_train_essay_tfidf_feat = tfidf_vectorizer(X_train,'essay',X_train)
x_test_essay_tfidf, x_test_essay_tfidf_feat = tfidf_vectorizer(X_train,'essay',X_test)
In [47]:
print(x_train_essay_tfidf.shape)
print(x_test_essay_tfidf.shape)
2.7.4) TFIDF (title)¶
In [48]:
from sklearn.feature_extraction.text import TfidfVectorizer
def tfidf_vectorizer_title(X_train,col_name,df):
vectorizer = TfidfVectorizer()
vectorizer.fit(X_train[col_name].values)
df_tfidf = vectorizer.transform(df[col_name].values)
return df_tfidf, vectorizer.get_feature_names()
In [49]:
# Lets vectorize essay
x_train_title_tfidf, x_train_title_tfidf_feat = tfidf_vectorizer_title(X_train,'project_title',X_train)
x_test_title_tfidf, x_test_title_tfidf_feat = tfidf_vectorizer_title(X_train,'project_title',X_test)
In [50]:
print(x_train_title_tfidf.shape)
print(x_test_title_tfidf.shape)
2.7.5) Using Pretrained Models: Avg W2V¶
In [51]:
'''
# Reading glove vectors in python: https://stackoverflow.com/a/38230349/4084039
def loadGloveModel(gloveFile):
print ("Loading Glove Model")
f = open(gloveFile,'r', encoding="utf8")
model = {}
for line in tqdm(f):
splitLine = line.split()
word = splitLine[0]
embedding = np.array([float(val) for val in splitLine[1:]])
model[word] = embedding
print ("Done.",len(model)," words loaded!")
return model
model = loadGloveModel('glove.42B.300d.txt')
# ============================
Output:
Loading Glove Model
1917495it [06:32, 4879.69it/s]
Done. 1917495 words loaded!
# ============================
words = []
for i in preproced_texts:
words.extend(i.split(' '))
for i in preproced_titles:
words.extend(i.split(' '))
print("all the words in the coupus", len(words))
words = set(words)
print("the unique words in the coupus", len(words))
inter_words = set(model.keys()).intersection(words)
print("The number of words that are present in both glove vectors and our coupus", \
len(inter_words),"(",np.round(len(inter_words)/len(words)*100,3),"%)")
words_courpus = {}
words_glove = set(model.keys())
for i in words:
if i in words_glove:
words_courpus[i] = model[i]
print("word 2 vec length", len(words_courpus))
# stronging variables into pickle files python: http://www.jessicayung.com/how-to-use-pickle-to-save-and-load-variables-in-python/
import pickle
with open('glove_vectors', 'wb') as f:
pickle.dump(words_courpus, f)
'''
Out[51]:
In [52]:
# stronging variables into pickle files python: http://www.jessicayung.com/how-to-use-pickle-to-save-and-load-variables-in-python/
# make sure you have the glove_vectors file
with open('glove_vectors', 'rb') as f:
model = pickle.load(f)
glove_words = set(model.keys())
In [53]:
# Combining all the above stundents
from tqdm import tqdm
def preprocess_essay(df,col_name):
preprocessed_essays = []
# tqdm is for printing the status bar
for sentance in tqdm(df[col_name].values):
sent = decontracted(sentance)
sent = sent.replace('\\r', ' ')
sent = sent.replace('\\"', ' ')
sent = sent.replace('\\n', ' ')
sent = re.sub('[^A-Za-z0-9]+', ' ', sent)
# https://gist.github.com/sebleier/554280
sent = ' '.join(e for e in sent.split() if e not in stopwords)
preprocessed_essays.append(sent.lower().strip())
return preprocessed_essays
In [54]:
# average Word2Vec
# compute average word2vec for each review.
def compute_avg_W2V(preprocessed_feature):
avg_w2v_vectors = []; # the avg-w2v for each sentence/review is stored in this list
for sentence in tqdm(preprocessed_feature): # for each review/sentence
vector = np.zeros(300) # as word vectors are of zero length
cnt_words =0; # num of words with a valid vector in the sentence/review
for word in sentence.split(): # for each word in a review/sentence
if word in glove_words:
vector += model[word]
cnt_words += 1
if cnt_words != 0:
vector /= cnt_words
avg_w2v_vectors.append(vector)
return avg_w2v_vectors
In [55]:
x_train_preprocessed_essay = preprocess_essay(X_train,'essay')
x_test_preprocessed_essay = preprocess_essay(X_test,'essay')
In [56]:
x_train_preprocessed_title = preprocess_essay(X_train,'project_title')
x_test_preprocessed_title = preprocess_essay(X_test,'project_title')
In [57]:
x_train_avg_w2v_essay = compute_avg_W2V(x_train_preprocessed_essay)
x_test_avg_w2v_essay = compute_avg_W2V(x_test_preprocessed_essay)
In [58]:
x_train_avg_w2v_title = compute_avg_W2V(x_train_preprocessed_title)
x_test_avg_w2v_title = compute_avg_W2V(x_test_preprocessed_title)
2.7.6) Using Pretrained Models: TFIDF Weighted W2V¶
In [59]:
# S = ["abc def pqr", "def def def abc", "pqr pqr def"]
def get_tfidf_dict(preprocessed_feature):
tfidf_model = TfidfVectorizer()
tfidf_model.fit(preprocessed_feature)
# we are converting a dictionary with word as a key, and the idf as a value
dictionary = dict(zip(tfidf_model.get_feature_names(), list(tfidf_model.idf_)))
tfidf_words = set(tfidf_model.get_feature_names())
return dictionary, tfidf_words
In [60]:
# average Word2Vec
# compute average word2vec for each review.
def compute_tfidf_w2v_vectors(preprocessed_feature):
tfidf_w2v_vectors = []; # the avg-w2v for each sentence/review is stored in this list
dictionary, tfidf_words = get_tfidf_dict(preprocessed_feature)
for sentence in tqdm(preprocessed_feature): # for each review/sentence
vector = np.zeros(300) # as word vectors are of zero length
tf_idf_weight =0; # num of words with a valid vector in the sentence/review
for word in sentence.split(): # for each word in a review/sentence
if (word in glove_words) and (word in tfidf_words):
vec = model[word] # getting the vector for each word
# here we are multiplying idf value(dictionary[word]) and the tf value((sentence.count(word)/len(sentence.split())))
tf_idf = dictionary[word]*(sentence.count(word)/len(sentence.split())) # getting the tfidf value for each word
vector += (vec * tf_idf) # calculating tfidf weighted w2v
tf_idf_weight += tf_idf
if tf_idf_weight != 0:
vector /= tf_idf_weight
tfidf_w2v_vectors.append(vector)
return tfidf_w2v_vectors
In [61]:
x_train_weighted_w2v_essay = compute_tfidf_w2v_vectors(x_train_essay_preprocessed)
x_test_weighted_w2v_essay= compute_tfidf_w2v_vectors(x_test_essay_preprocessed)
In [62]:
x_train_weighted_w2v_title = compute_tfidf_w2v_vectors(x_train_title_preprocessed)
x_test_weighted_w2v_title= compute_tfidf_w2v_vectors(x_test_title_preprocessed)
2.7.7) Vectorizing Numerical Features¶
We have 2 numerical features left, "price" and "teacher_number_of_previously_posted_projects". Let's check for the "missing" or "NaN" values present in those numerical features and use "Mean Replacement" for "price" and "Mode Replacement" for "teacher_number_of_previously_posted_projects".
In [63]:
print("Total number of \"Missing\" Values present in X_train price:",X_train['price'].isna().sum())
print("Total number of \"Missing\" Values present in X_test price:",X_test['price'].isna().sum())
In [64]:
print("Total number of \"Missing\" Values present in X_train previous teacher number:",X_train['teacher_number_of_previously_posted_projects'].isna().sum())
print("Total number of \"Missing\" Values present in X_test previous teacher number:",X_test['teacher_number_of_previously_posted_projects'].isna().sum())
In [65]:
print("Total number of \"Missing\" Values present in X_train quantity:",X_train['quantity'].isna().sum())
print("Total number of \"Missing\" Values present in X_test quantity:",X_test['quantity'].isna().sum())
"teacher_number_of_previously_posted_projects" does not have any "missing" values.
In [66]:
X_train['price'].mean()
Out[66]:
In [67]:
X_train['price'] = X_train['price'].fillna(287.1204)
In [68]:
X_test['price'].mean()
Out[68]:
In [69]:
X_test['price'] = X_test['price'].fillna(291.0478)
In [70]:
print(X_train['quantity'].mean())
print(X_test['quantity'].mean())
In [71]:
X_train['quantity'] = X_train['quantity'].fillna(19.1429)
X_test['quantity'] = X_test['quantity'].fillna(17.5885)
In [72]:
# check this one: https://www.youtube.com/watch?v=0HOqOcln3Z4&t=530s
# https://scikit-learn.org/stable/modules/generated/sklearn.preprocessing.StandardScaler.html
from sklearn.preprocessing import StandardScaler
def scaler_function(df,col_name):
scaler = StandardScaler()
scaler.fit(df[col_name].values.reshape(-1,1)) # finding the mean and standard deviation of this data
# Now standardize the data with above maen and variance.
print(f"Mean : {scaler.mean_[0]}, Standard deviation : {np.sqrt(scaler.var_[0])}")
scaled = scaler.transform(df[col_name].values.reshape(-1, 1))
return scaled
teacher_number_of_previously_posted_projects¶
In [73]:
x_train_teacher_number = scaler_function(X_train,'teacher_number_of_previously_posted_projects')
x_test_teacher_number = scaler_function(X_test,'teacher_number_of_previously_posted_projects')
price¶
In [74]:
x_train_price = scaler_function(X_train,'price')
x_test_price = scaler_function(X_test,'price')
In [75]:
x_train_quantity = scaler_function(X_train,'quantity')
x_test_quantity = scaler_function(X_test,'quantity')
2.8) Merging all the features and building the sets¶
In [76]:
# train dataset
print("After Vectorization and One hot encoding train dataset shape becomes:")
print(x_train_cat_one_hot.shape)
print(x_train_subcat_one_hot.shape)
print(x_train_state_one_hot.shape)
print(x_train_teacher_prefix_one_hot.shape)
print(x_train_grade_one_hot.shape)
print(x_train_essay_bow.shape)
print(x_train_title_bow.shape)
print(x_train_essay_tfidf.shape)
print(x_train_title_tfidf.shape)
print(np.asarray(x_train_avg_w2v_essay).shape)
print(np.asarray(x_train_avg_w2v_title).shape)
print(np.asarray(x_train_weighted_w2v_essay).shape)
print(np.asarray(x_train_weighted_w2v_title).shape)
print(x_train_teacher_number.shape)
print(x_train_price.shape)
print(x_train_quantity.shape)
print("="*50)
# test dataset
print("After Vectorization and One hot encoding test dataset shape becomes:")
print(x_test_cat_one_hot.shape)
print(x_test_subcat_one_hot.shape)
print(x_test_state_one_hot.shape)
print(x_test_teacher_prefix_one_hot.shape)
print(x_test_grade_one_hot.shape)
print(x_test_essay_bow.shape)
print(x_test_title_bow.shape)
print(x_test_essay_tfidf.shape)
print(x_test_title_tfidf.shape)
print(np.asarray(x_test_avg_w2v_essay).shape)
print(np.asarray(x_test_avg_w2v_title).shape)
print(np.asarray(x_test_weighted_w2v_essay).shape)
print(np.asarray(x_test_weighted_w2v_title).shape)
print(x_test_teacher_number.shape)
print(x_test_price.shape)
print(x_test_quantity.shape)
print("="*50)
2.8.1) Set 1: categorical, numerical features + project_title(BOW) + preprocessed_eassay (BOW with bi-grams with min_df=10 and max_features=5000)¶
In [77]:
# merge two sparse matrices: https://stackoverflow.com/a/19710648/4084039
from scipy.sparse import hstack
# with the same hstack function we are concatinating a sparse matrix and a dense matirx :)
X_train_set_1 = hstack((x_train_cat_one_hot,x_train_subcat_one_hot,x_train_state_one_hot,x_train_teacher_prefix_one_hot,\
x_train_grade_one_hot,x_train_teacher_number,x_train_price,x_train_quantity,x_train_title_bow,x_train_essay_bow)).tocsr()
X_test_set_1 = hstack((x_test_cat_one_hot,x_test_subcat_one_hot,x_test_state_one_hot,x_test_teacher_prefix_one_hot,\
x_test_grade_one_hot,x_test_teacher_number,x_test_price,x_test_quantity,x_test_title_bow,x_test_essay_bow)).tocsr()
In [78]:
# https://scikit-learn.org/stable/modules/generated/sklearn.model_selection.GridSearchCV.html
# for class prior i referred https://scikit-learn.org/stable/modules/generated/sklearn.naive_bayes.MultinomialNB.html,\
# and https://stackoverflow.com/questions/42498208/setting-prior-probabilities-in-naive-bayes-multinomialnb
from sklearn.model_selection import GridSearchCV
from scipy.stats import randint as sp_randint
from sklearn.model_selection import RandomizedSearchCV
import matplotlib.pyplot as plt
from sklearn.metrics import roc_auc_score
from sklearn.linear_model import LogisticRegression
import math
log_reg = LogisticRegression(class_weight = "balanced")
parameters = {'C':[10**x for x in range(-4,5)]}
clf = RandomizedSearchCV(log_reg, parameters,n_iter = 9, scoring='roc_auc')
clf.fit(X_train_set_1, y_train)
results = pd.DataFrame.from_dict(clf.cv_results_)
results = results.sort_values(['param_C'])
train_auc= results['mean_train_score']
train_auc_std= results['std_train_score']
cv_auc = results['mean_test_score']
cv_auc_std= results['std_test_score']
C_ = results['param_C'].apply(lambda x: math.log10(x))
plt.plot(C_, train_auc, label='Train AUC')
plt.plot(C_, cv_auc, label='CV AUC')
plt.scatter(C_, train_auc, label='Train AUC points')
plt.scatter(C_, cv_auc, label='CV AUC points')
plt.legend()
plt.xlabel("log10(C): hyperparameter")
plt.ylabel("AUC")
plt.title("Hyper parameter Vs AUC plot")
plt.grid()
plt.show()
results.head()
Out[78]:
In [79]:
# From the AUC plot, we find that the best value for "C" - "Inverse of Regularization Strength" for the LogisticRegression is 0.01
best_C = 0.001
In [80]:
# https://scikit-learn.org/stable/modules/generated/sklearn.metrics.roc_curve.html#sklearn.metrics.roc_curve
from sklearn.metrics import roc_curve, auc
log_reg = LogisticRegression(C=best_C, class_weight = "balanced")
log_reg.fit(X_train_set_1, y_train)
# roc_auc_score(y_true, y_score) the 2nd parameter should be probability estimates of the positive class
# not the predicted outputs
y_train_pred = log_reg.predict_proba(X_train_set_1)
y_test_pred = log_reg.predict_proba(X_test_set_1)
y_train_pred_prob = []
y_test_pred_prob = []
for index in range(len(y_train_pred)):
y_train_pred_prob.append(y_train_pred[index][1])
for index in range(len(y_test_pred)):
y_test_pred_prob.append(y_test_pred[index][1])
train_fpr, train_tpr, tr_thresholds = roc_curve(y_train, y_train_pred_prob)
test_fpr, test_tpr, te_thresholds = roc_curve(y_test, y_test_pred_prob)
plt.plot(train_fpr, train_tpr, label="train AUC ="+str(auc(train_fpr, train_tpr)))
plt.plot(test_fpr, test_tpr, label="test AUC ="+str(auc(test_fpr, test_tpr)))
plt.legend()
plt.xlabel("tpr")
plt.ylabel("fpr")
plt.title("ROC PLOTS for train,test")
plt.grid()
plt.show()
In [81]:
def compute_auc_with_hyper_para(x_tr,y_tr,x_cv,y_cv,para_list):
auc_tr = []
auc_cv = []
y_train_pred_prob = []
y_cv_pred_prob = []
for para in para_list:
log_reg = LogisticRegression(C = para, class_weight = "balanced")
log_reg.fit(x_tr,y_tr)
y_train_pred = log_reg.predict_proba(x_tr)
y_cv_pred = log_reg.predict_proba(x_cv)
for index in range(len(y_train_pred)):
y_train_pred_prob.append(y_train_pred[index][1])
for index in range(len(y_cv_pred)):
y_cv_pred_prob.append(y_cv_pred[index][1])
train_fpr, train_tpr, tr_thresholds = roc_curve(y_tr, y_train_pred_prob)
cv_fpr, cv_tpr, tc_thresholds = roc_curve(y_cv, y_cv_pred_prob)
y_train_pred_prob = []
y_cv_pred_prob = []
auc_tr.append(auc(train_fpr,train_tpr))
auc_cv.append(auc(cv_fpr,cv_tpr))
plt.plot(para_list,auc_tr,label="train_auc")
plt.plot(para_list,auc_cv,label="cv_auc")
plt.legend()
plt.xlabel("Hyper Parameter:C")
plt.ylabel("Area Under ROC Curve")
plt.title("auc v/s hyper parameters")
plt.grid()
plt.show()
In [82]:
para_list = [10**x for x in range(-4,5)]
compute_auc_with_hyper_para(X_train_set_1,y_train,X_test_set_1,y_test,para_list)
In [83]:
# we are writing our own function for predict, with defined thresould
# we will pick a threshold that will give the least fpr
def find_best_threshold(threshould, fpr, tpr):
t = threshould[np.argmax(tpr*(1-fpr))]
# (tpr*(1-fpr)) will be maximum if your fpr is very low and tpr is very high
print("the maximum value of tpr*(1-fpr)", max(tpr*(1-fpr)), "for threshold", np.round(t,3))
return t
def predict_with_best_t(proba, threshould):
predictions = []
for i in proba:
if i>=threshould:
predictions.append(1)
else:
predictions.append(0)
return predictions
In [84]:
print("="*100)
from sklearn.metrics import confusion_matrix
best_t = find_best_threshold(tr_thresholds, train_fpr, train_tpr)
print("Train confusion matrix")
sns.heatmap(confusion_matrix(y_train, predict_with_best_t(y_train_pred_prob, best_t)),annot = True, fmt = "d", cbar=False)
Out[84]:
In [85]:
print("Test confusion matrix")
sns.heatmap(confusion_matrix(y_test, predict_with_best_t(y_test_pred_prob, best_t)), annot = True, fmt = "d", cbar=False)
Out[85]:
2.8.2) Set 2: categorical, numerical features + project_title(TFIDF)+ preprocessed_eassay (TFIDF with bi-grams with min_df=10 and max_features=5000)¶
In [86]:
X_train_set_2 = hstack((x_train_cat_one_hot,x_train_subcat_one_hot,x_train_state_one_hot,x_train_teacher_prefix_one_hot,\
x_train_grade_one_hot,x_train_teacher_number,x_train_price,x_train_quantity,x_train_title_tfidf,x_train_essay_tfidf)).tocsr()
X_test_set_2 = hstack((x_test_cat_one_hot,x_test_subcat_one_hot,x_test_state_one_hot,x_test_teacher_prefix_one_hot,\
x_test_grade_one_hot,x_test_teacher_number,x_test_price,x_test_quantity,x_test_title_tfidf,x_test_essay_tfidf)).tocsr()
In [87]:
# https://scikit-learn.org/stable/modules/generated/sklearn.model_selection.GridSearchCV.html
# for class prior i referred https://scikit-learn.org/stable/modules/generated/sklearn.naive_bayes.MultinomialNB.html,\
# and https://stackoverflow.com/questions/42498208/setting-prior-probabilities-in-naive-bayes-multinomialnb
from sklearn.model_selection import GridSearchCV
from scipy.stats import randint as sp_randint
from sklearn.model_selection import RandomizedSearchCV
import matplotlib.pyplot as plt
from sklearn.metrics import roc_auc_score
from sklearn.linear_model import LogisticRegression
import math
log_reg = LogisticRegression(class_weight = "balanced")
parameters = {'C':[10**x for x in range(-4,5)]}
clf = RandomizedSearchCV(log_reg, parameters,n_iter = 9, scoring='roc_auc')
clf.fit(X_train_set_2, y_train)
results = pd.DataFrame.from_dict(clf.cv_results_)
results = results.sort_values(['param_C'])
train_auc= results['mean_train_score']
train_auc_std= results['std_train_score']
cv_auc = results['mean_test_score']
cv_auc_std= results['std_test_score']
C_ = results['param_C'].apply(lambda x: math.log10(x))
plt.plot(C_, train_auc, label='Train AUC')
plt.plot(C_, cv_auc, label='CV AUC')
plt.scatter(C_, train_auc, label='Train AUC points')
plt.scatter(C_, cv_auc, label='CV AUC points')
plt.legend()
plt.xlabel("log10(C): hyperparameter")
plt.ylabel("AUC")
plt.title("Hyper parameter Vs AUC plot")
plt.grid()
plt.show()
results.head()
Out[87]:
In [88]:
# From the AUC plot, we find that the best value for "C" - "Inverse of Regularization Strength" for the LogisticRegression is 0.01
best_C = 0.1
In [89]:
# https://scikit-learn.org/stable/modules/generated/sklearn.metrics.roc_curve.html#sklearn.metrics.roc_curve
from sklearn.metrics import roc_curve, auc
log_reg = LogisticRegression(C=best_C, class_weight = "balanced")
log_reg.fit(X_train_set_2, y_train)
y_train_pred = log_reg.predict_proba(X_train_set_2)
y_test_pred = log_reg.predict_proba(X_test_set_2)
y_train_pred_prob = []
y_test_pred_prob = []
for index in range(len(y_train_pred)):
y_train_pred_prob.append(y_train_pred[index][1])
for index in range(len(y_test_pred)):
y_test_pred_prob.append(y_test_pred[index][1])
train_fpr, train_tpr, tr_thresholds = roc_curve(y_train, y_train_pred_prob)
test_fpr, test_tpr, te_thresholds = roc_curve(y_test, y_test_pred_prob)
plt.plot(train_fpr, train_tpr, label="train AUC ="+str(auc(train_fpr, train_tpr)))
plt.plot(test_fpr, test_tpr, label="test AUC ="+str(auc(test_fpr, test_tpr)))
plt.legend()
plt.xlabel("tpr")
plt.ylabel("fpr")
plt.title("ROC PLOTS for train,test")
plt.grid()
plt.show()
In [90]:
para_list = [10**x for x in range(-4,5)]
compute_auc_with_hyper_para(X_train_set_2,y_train,X_test_set_2,y_test,para_list)
In [91]:
print("="*100)
from sklearn.metrics import confusion_matrix
best_t = find_best_threshold(tr_thresholds, train_fpr, train_tpr)
print("Train confusion matrix")
sns.heatmap(confusion_matrix(y_train, predict_with_best_t(y_train_pred_prob, best_t)),annot = True, fmt = "d", cbar=False)
Out[91]:
In [92]:
print("Test confusion matrix")
sns.heatmap(confusion_matrix(y_test, predict_with_best_t(y_test_pred_prob, best_t)), annot = True, fmt = "d", cbar=False)
Out[92]:
2.8.3) Set 3: categorical, numerical features + project_title(AVG W2V)+ preprocessed_eassay (AVG W2V)¶
In [93]:
X_train_set_3 = hstack((x_train_cat_one_hot,x_train_subcat_one_hot,x_train_state_one_hot,x_train_teacher_prefix_one_hot,\
x_train_grade_one_hot,x_train_teacher_number,x_train_price,x_train_quantity,x_train_avg_w2v_title,x_train_avg_w2v_essay)).tocsr()
X_test_set_3 = hstack((x_test_cat_one_hot,x_test_subcat_one_hot,x_test_state_one_hot,x_test_teacher_prefix_one_hot,\
x_test_grade_one_hot,x_test_teacher_number,x_test_price,x_test_quantity,x_test_avg_w2v_title,x_test_avg_w2v_essay)).tocsr()
In [94]:
# https://scikit-learn.org/stable/modules/generated/sklearn.model_selection.GridSearchCV.html
# for class prior i referred https://scikit-learn.org/stable/modules/generated/sklearn.naive_bayes.MultinomialNB.html,\
# and https://stackoverflow.com/questions/42498208/setting-prior-probabilities-in-naive-bayes-multinomialnb
from sklearn.model_selection import GridSearchCV
from scipy.stats import randint as sp_randint
from sklearn.model_selection import RandomizedSearchCV
import matplotlib.pyplot as plt
from sklearn.metrics import roc_auc_score
from sklearn.linear_model import LogisticRegression
import math
log_reg = LogisticRegression(class_weight = "balanced")
parameters = {'C':[10**x for x in range(-4,5)]}
clf = RandomizedSearchCV(log_reg, parameters,n_iter = 9, scoring='roc_auc')
clf.fit(X_train_set_3, y_train)
results = pd.DataFrame.from_dict(clf.cv_results_)
results = results.sort_values(['param_C'])
train_auc= results['mean_train_score']
train_auc_std= results['std_train_score']
cv_auc = results['mean_test_score']
cv_auc_std= results['std_test_score']
C_ = results['param_C'].apply(lambda x: math.log10(x))
plt.plot(C_, train_auc, label='Train AUC')
plt.plot(C_, cv_auc, label='CV AUC')
plt.scatter(C_, train_auc, label='Train AUC points')
plt.scatter(C_, cv_auc, label='CV AUC points')
plt.legend()
plt.xlabel("log10(C): hyperparameter")
plt.ylabel("AUC")
plt.title("Hyper parameter Vs AUC plot")
plt.grid()
plt.show()
results.head()
Out[94]:
In [95]:
# From the AUC plot, we find that the best value for "C" - "Inverse of Regularization Strength" for the LogisticRegression is 0.01
best_C = 1
In [96]:
# https://scikit-learn.org/stable/modules/generated/sklearn.metrics.roc_curve.html#sklearn.metrics.roc_curve
from sklearn.metrics import roc_curve, auc
log_reg = LogisticRegression(C=best_C, class_weight = "balanced")
log_reg.fit(X_train_set_3, y_train)
y_train_pred = log_reg.predict_proba(X_train_set_3)
y_test_pred = log_reg.predict_proba(X_test_set_3)
y_train_pred_prob = []
y_test_pred_prob = []
for index in range(len(y_train_pred)):
y_train_pred_prob.append(y_train_pred[index][1])
for index in range(len(y_test_pred)):
y_test_pred_prob.append(y_test_pred[index][1])
train_fpr, train_tpr, tr_thresholds = roc_curve(y_train, y_train_pred_prob)
test_fpr, test_tpr, te_thresholds = roc_curve(y_test, y_test_pred_prob)
plt.plot(train_fpr, train_tpr, label="train AUC ="+str(auc(train_fpr, train_tpr)))
plt.plot(test_fpr, test_tpr, label="test AUC ="+str(auc(test_fpr, test_tpr)))
plt.legend()
plt.xlabel("tpr")
plt.ylabel("fpr")
plt.title("ROC PLOTS for train,test")
plt.grid()
plt.show()
In [98]:
para_list = [10**x for x in range(-4,5)]
compute_auc_with_hyper_para(X_train_set_3,y_train,X_test_set_3,y_test,para_list)
In [99]:
print("="*100)
from sklearn.metrics import confusion_matrix
best_t = find_best_threshold(tr_thresholds, train_fpr, train_tpr)
print("Train confusion matrix")
sns.heatmap(confusion_matrix(y_train, predict_with_best_t(y_train_pred_prob, best_t)),annot = True, fmt = "d", cbar=False)
Out[99]:
In [100]:
print("Test confusion matrix")
sns.heatmap(confusion_matrix(y_test, predict_with_best_t(y_test_pred_prob, best_t)), annot = True, fmt = "d", cbar=False)
Out[100]:
2.8.4) Set 4: categorical, numerical features + project_title(TFIDF W2V)+ preprocessed_essay (TFIDF W2V)¶
In [101]:
X_train_set_4 = hstack((x_train_cat_one_hot,x_train_subcat_one_hot,x_train_state_one_hot,x_train_teacher_prefix_one_hot,\
x_train_grade_one_hot,x_train_teacher_number,x_train_price,x_train_quantity,x_train_weighted_w2v_title,x_train_weighted_w2v_essay)).tocsr()
X_test_set_4 = hstack((x_test_cat_one_hot,x_test_subcat_one_hot,x_test_state_one_hot,x_test_teacher_prefix_one_hot,\
x_test_grade_one_hot,x_test_teacher_number,x_test_price,x_test_quantity,x_test_weighted_w2v_title,x_test_weighted_w2v_essay)).tocsr()
In [102]:
# https://scikit-learn.org/stable/modules/generated/sklearn.model_selection.GridSearchCV.html
# for class prior i referred https://scikit-learn.org/stable/modules/generated/sklearn.naive_bayes.MultinomialNB.html,\
# and https://stackoverflow.com/questions/42498208/setting-prior-probabilities-in-naive-bayes-multinomialnb
from sklearn.model_selection import GridSearchCV
from scipy.stats import randint as sp_randint
from sklearn.model_selection import RandomizedSearchCV
import matplotlib.pyplot as plt
from sklearn.metrics import roc_auc_score
from sklearn.linear_model import LogisticRegression
import math
log_reg = LogisticRegression()
parameters = {'C':[10**x for x in range(-4,5)]}
clf = RandomizedSearchCV(log_reg, parameters,n_iter = 9, scoring='roc_auc')
clf.fit(X_train_set_4, y_train)
results = pd.DataFrame.from_dict(clf.cv_results_)
results = results.sort_values(['param_C'])
train_auc= results['mean_train_score']
train_auc_std= results['std_train_score']
cv_auc = results['mean_test_score']
cv_auc_std= results['std_test_score']
C_ = results['param_C'].apply(lambda x: math.log10(x))
plt.plot(C_, train_auc, label='Train AUC')
# this code is copied from here: https://stackoverflow.com/a/48803361/4084039
# plt.gca().fill_between(K, train_auc - train_auc_std,train_auc + train_auc_std,alpha=0.2,color='darkblue')
plt.plot(C_, cv_auc, label='CV AUC')
# # this code is copied from here: https://stackoverflow.com/a/48803361/4084039
# # plt.gca().fill_between(K, cv_auc - cv_auc_std,cv_auc + cv_auc_std,alpha=0.2,color='darkorange')
plt.scatter(C_, train_auc, label='Train AUC points')
plt.scatter(C_, cv_auc, label='CV AUC points')
plt.legend()
plt.xlabel("log10(C): hyperparameter")
plt.ylabel("AUC")
plt.title("Hyper parameter Vs AUC plot")
plt.grid()
plt.show()
results.head()
Out[102]:
In [103]:
# From the AUC plot, we find that the best value for "C" - "Inverse of Regularization Strength" for the LogisticRegression is 0.01
best_C = 0.1
In [104]:
# https://scikit-learn.org/stable/modules/generated/sklearn.metrics.roc_curve.html#sklearn.metrics.roc_curve
from sklearn.metrics import roc_curve, auc
log_reg = LogisticRegression(C=best_C, class_weight = "balanced")
log_reg.fit(X_train_set_4, y_train)
y_train_pred = log_reg.predict_proba(X_train_set_4)
y_test_pred = log_reg.predict_proba(X_test_set_4)
y_train_pred_prob = []
y_test_pred_prob = []
for index in range(len(y_train_pred)):
y_train_pred_prob.append(y_train_pred[index][1])
for index in range(len(y_test_pred)):
y_test_pred_prob.append(y_test_pred[index][1])
train_fpr, train_tpr, tr_thresholds = roc_curve(y_train, y_train_pred_prob)
test_fpr, test_tpr, te_thresholds = roc_curve(y_test, y_test_pred_prob)
plt.plot(train_fpr, train_tpr, label="train AUC ="+str(auc(train_fpr, train_tpr)))
plt.plot(test_fpr, test_tpr, label="test AUC ="+str(auc(test_fpr, test_tpr)))
plt.legend()
plt.xlabel("tpr")
plt.ylabel("fpr")
plt.title("ROC PLOTS for train,test")
plt.grid()
plt.show()
In [105]:
para_list = [10**x for x in range(-4,5)]
compute_auc_with_hyper_para(X_train_set_4,y_train,X_test_set_4,y_test,para_list)
In [106]:
print("="*100)
from sklearn.metrics import confusion_matrix
best_t = find_best_threshold(tr_thresholds, train_fpr, train_tpr)
print("Train confusion matrix")
sns.heatmap(confusion_matrix(y_train, predict_with_best_t(y_train_pred_prob, best_t)),annot = True, fmt = "d", cbar=False)
Out[106]:
In [107]:
print("Test confusion matrix")
sns.heatmap(confusion_matrix(y_test, predict_with_best_t(y_test_pred_prob, best_t)), annot = True, fmt = "d", cbar=False)
Out[107]:
2.8.5) Calculate Sentiment Score for each essay (combined)¶
In [112]:
import nltk
from nltk.sentiment.vader import SentimentIntensityAnalyzer
# import nltk
# nltk.download('vader_lexicon')
def compute_sentiment_score(df):
score_list = []
sid = SentimentIntensityAnalyzer()
for essay in df['essay']:
ss = sid.polarity_scores(essay)
score_list.append(ss)
return score_list
# we can use these 4 things as features/attributes (neg, neu, pos, compound)
# neg: 0.0, neu: 0.753, pos: 0.247, compound: 0.93
In [113]:
x_train_score = compute_sentiment_score(X_train)
x_test_score = compute_sentiment_score(X_test)
In [114]:
def populate_list(score_dicts):
neg_score = []
neu_score = []
pos_score = []
compound_score = []
for dict_ in score_dicts:
neg_score.append(dict_['neg'])
neu_score.append(dict_['neu'])
pos_score.append(dict_['pos'])
compound_score.append(dict_['compound'])
return neg_score, neu_score, pos_score, compound_score
In [115]:
x_train_neg, x_train_neu, x_train_pos, x_train_compound = populate_list(x_train_score)
x_test_neg, x_test_neu, x_test_pos, x_test_compound = populate_list(x_test_score)
In [116]:
X_train['words_project_title'] = X_train['project_title'].apply(lambda x: len(x.split()))
X_train['words_essay'] = X_train['essay'].apply(lambda x: len(x.split()))
In [117]:
X_test['words_project_title'] = X_test['project_title'].apply(lambda x: len(x.split()))
X_test['words_essay'] = X_test['essay'].apply(lambda x: len(x.split()))
In [119]:
# X_cv['words_project_title'] = X_cv['project_title'].apply(lambda x: len(x.split()))
# X_cv['words_essay'] = X_cv['essay'].apply(lambda x: len(x.split()))
Let's join the all the sentiment scores to the respective dataframes
In [120]:
# for training set
X_train['neg'] = x_train_neg
X_train['neu'] = x_train_neu
X_train['pos'] = x_train_pos
X_train['compound'] = x_train_compound
# for testing set
X_test['neg'] = x_test_neg
X_test['neu'] = x_test_neu
X_test['pos'] = x_test_pos
X_test['compound'] = x_test_compound
# # for cv set
# X_cv['neg'] = x_cv_neg
# X_cv['neu'] = x_cv_neu
# X_cv['pos'] = x_cv_pos
# X_cv['compound'] = x_cv_compound
In [121]:
x_train_neg_reshaped = X_train.values.reshape(-1,1)
2.8.6) Vectorizing newly added numerical features: words_project_title, words_essay¶
In [122]:
#Let's do for all the title words
x_train_words_title_scaled = scaler_function(X_train,'words_project_title')
x_test_words_title_scaled = scaler_function(X_test,'words_project_title')
# x_cv_words_title_scaled = scaler_function(X_cv,'words_project_title')
# let's do for all the essay words
x_train_words_essay_scaled = scaler_function(X_train,'words_essay')
x_test_words_essay_scaled = scaler_function(X_test,'words_essay')
# x_cv_words_essay_scaled = scaler_function(X_cv,'words_essay')
2.8.6) Prepare dataset 5¶
In [123]:
X_train_set_5 = hstack((x_train_cat_one_hot,x_train_subcat_one_hot,x_train_state_one_hot,x_train_teacher_prefix_one_hot,\
x_train_grade_one_hot,x_train_teacher_number,x_train_price,x_train_quantity,x_train_words_title_scaled,x_train_words_essay_scaled, X_train['neg'].values.reshape(-1,1),X_train['neu'].values.reshape(-1,1),X_train['pos'].values.reshape(-1,1),X_train['compound'].values.reshape(-1,1))).tocsr()
X_test_set_5 = hstack((x_test_cat_one_hot,x_test_subcat_one_hot,x_test_state_one_hot,x_test_teacher_prefix_one_hot,\
x_test_grade_one_hot,x_test_teacher_number,x_test_price,x_test_quantity,x_test_words_title_scaled,x_test_words_essay_scaled,X_test['neg'].values.reshape(-1,1),X_test['neu'].values.reshape(-1,1),X_test['pos'].values.reshape(-1,1),X_test['compound'].values.reshape(-1,1))).tocsr()
In [124]:
# https://scikit-learn.org/stable/modules/generated/sklearn.model_selection.GridSearchCV.html
# for class prior i referred https://scikit-learn.org/stable/modules/generated/sklearn.naive_bayes.MultinomialNB.html,\
# and https://stackoverflow.com/questions/42498208/setting-prior-probabilities-in-naive-bayes-multinomialnb
from sklearn.model_selection import GridSearchCV
from scipy.stats import randint as sp_randint
from sklearn.model_selection import RandomizedSearchCV
import matplotlib.pyplot as plt
from sklearn.metrics import roc_auc_score
from sklearn.linear_model import LogisticRegression
import math
log_reg = LogisticRegression(class_weight = "balanced")
parameters = {'C':[10**x for x in range(-4,5)]}
clf = RandomizedSearchCV(log_reg, parameters,n_iter = 9, scoring='roc_auc')
clf.fit(X_train_set_5, y_train)
results = pd.DataFrame.from_dict(clf.cv_results_)
results = results.sort_values(['param_C'])
train_auc= results['mean_train_score']
train_auc_std= results['std_train_score']
cv_auc = results['mean_test_score']
cv_auc_std= results['std_test_score']
C_ = results['param_C'].apply(lambda x: math.log10(x))
plt.plot(C_, train_auc, label='Train AUC')
# this code is copied from here: https://stackoverflow.com/a/48803361/4084039
# plt.gca().fill_between(K, train_auc - train_auc_std,train_auc + train_auc_std,alpha=0.2,color='darkblue')
plt.plot(C_, cv_auc, label='CV AUC')
# # this code is copied from here: https://stackoverflow.com/a/48803361/4084039
# # plt.gca().fill_between(K, cv_auc - cv_auc_std,cv_auc + cv_auc_std,alpha=0.2,color='darkorange')
plt.scatter(C_, train_auc, label='Train AUC points')
plt.scatter(C_, cv_auc, label='CV AUC points')
plt.legend()
plt.xlabel("log10(C): hyperparameter")
plt.ylabel("AUC")
plt.title("Hyper parameter Vs AUC plot")
plt.grid()
plt.show()
results.head()
Out[124]:
In [125]:
# From the AUC plot, we find that the best value for "C" - "Inverse of Regularization Strength" for the LogisticRegression is 0.01
best_C = 0.01
In [126]:
# https://scikit-learn.org/stable/modules/generated/sklearn.metrics.roc_curve.html#sklearn.metrics.roc_curve
from sklearn.metrics import roc_curve, auc
log_reg = LogisticRegression(C=best_C, class_weight = "balanced")
log_reg.fit(X_train_set_5, y_train)
y_train_pred = log_reg.predict_proba(X_train_set_5)
y_test_pred = log_reg.predict_proba(X_test_set_5)
y_train_pred_prob = []
y_test_pred_prob = []
for index in range(len(y_train_pred)):
y_train_pred_prob.append(y_train_pred[index][1])
for index in range(len(y_test_pred)):
y_test_pred_prob.append(y_test_pred[index][1])
train_fpr, train_tpr, tr_thresholds = roc_curve(y_train, y_train_pred_prob)
test_fpr, test_tpr, te_thresholds = roc_curve(y_test, y_test_pred_prob)
plt.plot(train_fpr, train_tpr, label="train AUC ="+str(auc(train_fpr, train_tpr)))
plt.plot(test_fpr, test_tpr, label="test AUC ="+str(auc(test_fpr, test_tpr)))
plt.legend()
plt.xlabel("tpr")
plt.ylabel("fpr")
plt.title("ROC PLOTS for train,test")
plt.grid()
plt.show()
In [127]:
para_list = [10**x for x in range(-4,5)]
compute_auc_with_hyper_para(X_train_set_5,y_train,X_test_set_5,y_test,para_list)
In [128]:
print("="*100)
from sklearn.metrics import confusion_matrix
best_t = find_best_threshold(tr_thresholds, train_fpr, train_tpr)
print("Train confusion matrix")
sns.heatmap(confusion_matrix(y_train, predict_with_best_t(y_train_pred_prob, best_t)),annot = True, fmt = "d", cbar=False)
Out[128]:
In [129]:
print("Test confusion matrix")
sns.heatmap(confusion_matrix(y_test, predict_with_best_t(y_test_pred_prob, best_t)), annot = True, fmt = "d", cbar=False)
Out[129]:
